Logo of nihpaAbout Author manuscriptsSubmit a manuscriptNIH Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
J Immunol. Author manuscript; available in PMC Oct 15, 2009.
Published in final edited form as:
PMCID: PMC2677919

Alterations of the medullary epithelial compartment in the Aire-deficient thymus: Implications for programs of thymic epithelial differentiation1


A widely held model of thymic epithelial differentiation is based on patterns of keratin expression, where a K8+K5+ progenitor gives rise to K8+K5/K14 cortical thymic epithelium (CTEC), and medullary thymic epithelium (MTEC) is K8K5+K14+. The thymic phenotype of p63-deficient mice indicates that p63 is an important regulator of proximal stages of TE differentiation. In this study we have examined several features of the thymic medullary compartment in WT and Aire-deficient thymi in an effort to integrate the pro-apoptotic activity of Aire with these different perspectives of TE differentiation. Patterns of keratin and p63 expression by MTEC described here are difficult to reconcile with postmitotic MTEC that express a K8K14+ phenotype and suggest that the patterns of p63 and keratin expression reflecting differentiation programs of other epithelial tissues provide a useful framework for revising models of TE differentiation. Alterations of the Aire−/− MTEC compartment included reduced expression of p63, increased frequency of MTEC expressing truncated Aire protein, and shifts in the pattern of keratin expression and epithelial morphology. These data suggest a scenario where cellular targets of Aire-mediated apoptosis are postmitotic MTEC that have not yet completed their terminal differentiation program. According to this view, the minor population of globular K8+K14−/low MTEC observed in the Aire+/+ thymus and that are significantly expanded in the Aire−/− thymic medulla represent end-stage, terminally differentiated MTEC. These Aire-dependent alterations of the MTEC compartment suggest that the activity of Aire is not neutral with respect to the program of MTEC differentiation.

Keywords: Thymus, Aire, p63, claudin, differentiation


Thymic epithelium contributes to an environment that recruits lymphoid progenitor cells from the circulation, directs a number of important fate choices during thymocyte development, and shapes the repertoire of T cell antigen diversity by imposing MHC restriction during positive selection. Thymic epithelial cells (TEC) also project a spectrum of self-antigens that contribute to negative selection of thymocytes with auto-reactive specificity. Based on anatomic, phenotypic, and functional criteria, thymic epithelium has been compartmentalized into cortical and medullary compartments, where cortical epithelium is held to be primarily responsible for directing positive selection and medullary epithelium is involved in the negative selection process (reviewed in (1, 2)).

The developmental program that gives rise to this epithelial heterogeneity remains poorly understood. Progenitor epithelial cells able to give rise to both cortical and medullary TEC have been demonstrated in the fetal (3) and adult thymus (4), although the phenotypic properties of these cells are not known (5). It is widely held that TEC co-expressing K8 and K5 represent precursors to TEC destined to be K8+K5/K14-cortical TEC (CTEC) (6, 7) and that TEC bearing a K5/K14+K8 phenotype define the majority of medullary TEC (MTEC). The differentiation program of medullary TE remains obscure.

Recently p63 has been implicated in TE differentiation by promoting the survival of progenitor/transit amplifying cells (8, 9). Some of this effect may reflect alteration of FGFR2IIIb signaling, since FGFR2IIIb is a downstream target of p63 (8), FGFR2IIIb signaling can affect levels of p63 expression (10), and the thymic phenotypes of FGFR2-deficient and p63 deficient thymi are quite similar (8, 9, 11, 12).

p63 has long been considered as a marker of basal progenitor/transit amplifying cells in stratified or pseudostratifed epithelia of epidermis, prostate, trachea and esophagus (9, 1317). In the epidermis, down-regulation of p63 expression in suprabasal cells is associated with cessation of cell division and initiation of terminal differentiation, with additional differentiation occurring within this postmitotic population (9, 13).

A third view of MTEC differentiation has emerged from the patterns of Aire expression among MTEC populations defined by cell surface phenotype, where Aire is preferentially expressed by MTEC bearing a CD80hi MHCIIhi phenotype. This phenotype is considered to reflect terminally differentiated MTEC based on a similar CD80/MHCII phenotype displayed by differentiated dendritic cells (18) and the observation that Aire+ MTEC are postmitotic and display a high degree of turnover (19). Transfection studies with TE cell lines suggested that Aire expression compromises cell survival and enhances apoptosis (19). Pleiomorphic alterations of the medullary thymic compartment as a consequence Aire-deficiency has also led to the suggestion that Aire may play a broader role in MTEC differentiation (20).

In an attempt to integrate these different views of TE differentiation and the possible role of Aire in this process, we have related Aire and p63 expression with the expression of keratins considered to define stages of TE differentiation in Aire+/+ and Aire−/− thymi. The data presented here raise a number of questions regarding some of the current views regarding the process of TE differentiation in general and MTEC differentiation in particular. First, extensive co-expression of K8 and K14 by MTEC suggest that the current binary model of K8+K5/K14 CTEC and K8K5/K14+ MTEC differentiation should be revised. We also demonstrate that MTEC in the Aire−/− thymus display altered morphology, altered patterns of p63 and keratin expression, and increased frequencies of MTEC that expressed non-functional Aire protein and may reflect MTEC that would be eliminated in the Aire+/+ thymus. While some of these alterations are consistent with the interpretation that Aire is involved in the elimination of end-stage, terminally differentiated MTEC, other alterations of the Aire−/− medullary compartment can be interpreted to indicate that Aire exerts pro-apoptotic effects at a more proximal stage of MTEC differentiation, and eliminates post-mitotic MTEC that have not completed their terminal differentiation program.

Materials and Methods


C57Bl/6 mice were obtained from Charles Rivers. One strain of Aire-deficient mice (21) was obtained from Jackson Laboratories (Bar Harbor, ME) and the other (22) was obtained from Dr. L. Peltonen (Department of Molecular Medicine, National Public Health Institute, Helsinki, Finland). All mice were maintained in the University of Washington Specific Pathogen Free facility and used in accordance with protocols approved by the University of Washington Institutional Animal Care and Use Committee.


Primary antibodies for immunohistochemistry included anti-EpCAM (G8.8;(23)), anti-podoplanin (8.1.1;(24)), an antibody that detects a 110 kDa protein expressed by medullary TE (10.1.1; (25)), and a rat IgM MAB that reacts with keratin 14 (3G10; (26)). G8.8 and 8.1.1 are available from the Developmental Studies Hybridoma Bank (http://dshb.biology.uiowa.edu/)). Two monoclonal antibodies preferentially reacting with cortical thymic epithelium were used; anti-DEC205 (NLDC145; (27) and CDR1 (28). Polyclonal anti-Aire antibodies (D-17; SC-17986 and M-300; SC-33189) and a polyclonal anti-p63 (H-137, SC-8343) were from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). Polyclonal rabbit anti-K5 and K14 antibodies were purchased from Covance (Princeton, NJ); monoclonal anti-K8 (Troma-1) (29) was obtained from the Developmental Studies Hybridoma Bank (http://dshb.biology.uiowa.edu/), and anti-claudin antibodies were purchased from Zymed Laboratories (San Francisco, CA). Secondary reagents for immunofluorescence microscopy (donkey anti-goat IgG, donkey anti-rabbit IgG, goat anti-rabbit IgG, goat anti-rat IgG, chicken anti-rat IgG and streptavidin conjugated with Alexa 488, Alexa 546, or Alexa 647) were purchased from Molecular Probes (Carlsbad, CA).


Different tissue processing protocols were used to accommodate the requirements of the reagent antibodies employed. Wherever possible, aldehyde fixation was used to minimize the diffusion artifacts introduced by acetone fixation. Frozen sections collected on microscope slides were fixed in 0.1M cacodylate buffer, pH7.4, containing 4% paraformaldehyde and 5 mM CaCl2. In some instances, thymus tissue was perfused via the heart with 10 ml of 1% paraformaldehyde fixative or immersed in this fixative for 4–16 hours @4°C, then washed repeatedly in PBS prior to cryoprotection (30% sucrose in PBS) and embedment in OCT (Sakura Finetech, Torrance, CA) for cryosectioning. Antigen retrieval was accomplished by incubating paraformaldehyde-fixed sections in 20 mM tris buffer, pH 9.0 for 25 minutes in a steam device (Black&Decker, HS2776, Towson, MD) and then allowing the sections to cool for ~45 minutes before washing in PBS and subsequent processing. To detect antigens sensitive to aldehyde fixation, freshly isolated thymic tissue was embedded in OCT, and sections mounted on glass slides were fixed in cold (−20°C) acetone for 20 minutes. Monoclonal antibodies were used as hybridoma supernatants, while polyclonal antibodies were used at dilutions of 1:500–2000. Best results with the anti-Aire and anti-p63 antibodies were obtained with aldehyde fixation prior to tissue sectioning and the antigen retrieval protocol described above.

Aldehyde-fixed sections were incubated in PBS containing 10% vol/vol of normal serum for 1 hour prior to application of primary antibodies (polyclonal antibodies diluted in hybridoma supernatant or PBS containing 1% (wt/vol) BSA and 10% normal serum). After incubation with primary antibodies overnight, sections were washed repeatedly with PBS and then incubated with appropriate secondary antibodies. After 1–1.5 hour incubation, sections were again washed and then coverslips were applied with Fluoromount G (Southern Biotech, Birmingham, AL). Processing of acetone-fixed samples was similar, except that primary antibody incubation times were reduced to 1 hour.

Images were captured with a fluorescence microscope equipped with a monochrome digital CCD camera (Orca-ER, Hamamatsu, Bridgewater, NJ ) and assembled into RGB images with Photoshop (Adobe, San Jose, CA). Determination of tissue areas examined and quantitation of labeled cells was accomplished with Image J (http://rsb.info.nih.gov/ij/). The Wilcoxon two-sample test was used to determine statistical significance of the resulting data http://www.fon.hum.uva.nl/Service/Statistics/Wilcoxon_Test.html

Detection of apoptotic cells

Tunel assay was performed as described (30), using streptavidin-Alexa 555 for detection of incorporated biotinylated nucleotides. Labeling procedures in the absence of TdT served as a negative control. A fluorescent pan-caspase inhibitor (cat.# G7461, Promega, Madison WI) was mixed with primary antibodies at a final concentration of 2.5 µM, using tissue sections of thymus fixed by perfusion with 1% paraformaldehyde.

PCR analyses

Procedures for enzymatic dissociation of thymic tissue, flow cytometric purification of MTEC, obtention of RNA from these cells and generation of cDNA have been described previously (12, 20, 31). Primer sequences are available upon request. Determination of RQ values, using three independently generated pools of sorted MTEC cDNA has been described previously (12, 20, 31) and standard deviations of real time data were calculated according to manufacturer’s suggestions (www.appliedbiosystems.com).


Detection of Aire protein in WT and Aire-deficient thymus

Recent data indicating that the MTEC expressing Aire are post-mitotic and display a high degree of turnover led to the suggestion that Aire expression contributes to the relatively short half-life of Aire+ MTEC (19). The MTEC eliminated by Aire activity may be an end stage, terminally differentiated population or may correspond to MTEC that are postmitotic, but have not completed their terminal differentiation program. In the latter case, the absence of functional Aire protein would reveal the full program of MTEC terminal differentiation. We first evaluated the ability of commercially available anti-Aire antibodies to detect the expression of truncated Aire protein expressed by two independently generated lines of Aire−/− mice as a means to identify MTEC that would normally be eliminated by Aire activity. One line has a Cre-mediated deletion in exon 1–2 (21), while the other carries a disruption of exons 5&6 of the Aire gene (22).

As shown in Figure 1a, two polyclonal anti-Aire antibodies reacted with a subset of medullary epithelial cells of Aire+/+ thymi with the speckled and punctate staining patteren described previously with other anti-Aire antibodies (3234). The M300 antibody was raised in rabbits against a peptide fragment (253–552) mapping at the C-terminus of mouse Aire, while the D17 antibody was raised in goats against an internal peptide of the human AIRE protein. The staining pattern of D17 and M300 were totally overlapping in the normal thymus, indicating that they were detected epitopes associated with a common structure. As expected, thymic epithelial cells from mutant mice expressing an exon 1–2 truncation of Aire failed to react with either the M300 or D17 antibodies (Figure 1b, middle column). The M300 antibody also did not react with thymic tissue from Aire−/− mice bearing a deletion of exons 5 and 6 of the Aire gene and also lacking the carboxy-terminal Aire epitope. However, the D17 antibody did detect the truncated Aire protein produced by the Peltonen strain of Aire-deficient mice (Figure 1b, right column). The pattern of D17 staining of MTEC in this strain of Aire−/− mice was sensitive to tissue preparation, while the pattern of Aire detection in the Aire+/+ thymus was not. When sections of fresh frozen tissue were fixed with cold acetone or paraformaldehyde, staining with D17 usually presented as one or two punctate dots over the nucleus, with some diffuse nuclear and perhaps cytoplasmic staining. In contrast, tissue fixation with paraformaldehyde prior to freezing and sectioning yielded an intense, but diffuse nuclear staining pattern in addition to the variable cytoplasmic staining. Because fixation prior to tissue manipulation minimizes the introduction of preparation artifacts and can provide material suitable for ultrastructural studies (2325), the pattern of Aire staining seen in “prefixed” tissue is likely to more accurately reflect the distribution of Aire protein. The diffuse nuclear localization of the truncated Aire protein observed in the Peltonen Aire−/− mice presumably reflects the retention of amino-terminal nuclear localization signals and the absence of Aire domains that interact with DNA.

Figure 1
Detection of Aire protein in Aire+/+ and Aire−/− thymus. a. Upper row of panels demonstrate that two polyclonal anti-Aire antibodies, D17 (goat anti-Aire) and M300 (rabbit anti-Aire) yield the same speckled and punctate staining pattern ...

Patterns of Aire expression by MTEC in Aire+/+ and Aire−/− thymi

We used the D17 antibody in conjunction with other anti-stromal cell antibodies to better characterize the phenotype of MTEC that express Aire protein in Aire+/+ and Aire−/− thymi. As shown in Figure 2a–c, Aire expression was preferentially associated with K14+K8+ MTEC in the Aire+/+ thymus, with rare K8+K14 MTEC also expressing Aire. As described previously (12, 20) and shown here, K14+K8+ cells comprises a prominent MTEC subset in the Aire+/+ thymus and display a range of K14/K8 expression ratios, making it difficult to relate these findings to reports that the majority of MTEC express a K5/K14+K8 phenotype (6, 7). We had previously demonstrated that the medullary compartment of Aire−/− mice had an elevated representation of K8+ globular MTEC with variably low expression of K5 or K14 (20). Some of these globular MTEC in the Aire−/− thymus were D17+, indicating that they may be derived from K14+K8+ MTEC and thus may represent cells that would normally express Aire and thus normally be eliminated in the Aire+/+ thymus. The increased frequency of globular K8+, variable K14+ MTEC in the Aire−/− medulla was accompanied by a coarser reticular MTEC arrangement, with fewer stellate-shaped cells (compare panels 2a–c with 2d–e). Since it is not known if Aire expression, once induced, is constitutively expressed, or if Aire expression is associated with a discrete stage of MTEC differentiation, the mixture of Aire+ and Aire MTEC in the Aire−/− medulla may represent progeny of Aire+ and Aire precursors or may indicate that surviving progeny of Aire+ precursors acquire an Aire phenotype as they undergo additional differentiation.

Figure 2
Aire expression associated with MTEC populations defined by K8/K14 expression. Samples of young adult (a–f) or neonatal (g–l) thymus from Aire+/+ (a–c, g–i) and Aire−/− (d–f, j–l) were processed ...

A similar situation was seen in neonatal thymic tissue, where Aire expression was preferentially associated with K8+K14+ MTEC and a few K8+K14−/low cells (compare panels 2g–i with 2j–l). The accumulation of Aire+ cells in the Aire−/− neonatal thymus was less pronounced compared to thymi of older mice, suggesting that this phenotype is acquired over time, consistent with the notion that the accumulation of these cells is due to altered developmental kinetics rather than direct Aire activity.

The relative density of D17+ cells in Aire +/+ and Aire−/− thymus tissue shown in Figure 2, indicated a higher density of Aire+ cells in the Aire−/− thymus. Enumeration of D17+ cells in 3 sets of Aire−/− and Aire +/+ thymi confirmed this and demonstrated that this difference was statistically significant (Figure 3).

Figure 3
The frequency of Aire+ cells is increased in the Aire−/− thymus. Multiple thymus sections from three individual mice were used to quantitate the number of Aire+ cells within the medullary compartment defined by K8 and K14 expression.

Aire deficiency alters p63 expression in the medullary epithelial compartment

It has recently been reported that p63 plays an important role in maintaining progenitor epithelial cells in the thymus (9), perhaps in part through the regulation of FGFR2IIIb and Jagged expression by thymic epithelium (8). Because of the general alterations of the Aire-deficient medullary compartment described here and previously (20), we evaluated the impact of Aire deficiency on p63 expression by TE, reasoning that Aire could affect other pathways regulating MTEC differentiation. As shown in Figure 4, and in agreement with previous reports (3538), p63 expression was widespread among epithelial cells in cortical and medullary compartments, with a higher density of p63+ cells in the medulla (Figure 4a). Within the medullary compartment, there was little overlap between p63 expression and either Aire or high levels of K8 (Figure 4b & c). In contrast, there was good concordance between K14 and p63 expression and between K14 and Aire expression (Figure 4g&h), indicating that p63 and Aire are expressed by different subsets of MTEC or by a common population of MTEC at different stages of their differentiation. In the Aire−/− thymus the density of medullary p63+ cells was noticeably reduced, again with few cells expressing both Aire and p63, similar to the Aire+/+ thymus (Figure 4b). The enhanced expression of truncated Aire protein by K8+ MTEC described above is also evident in Figure 4 e&i. A lower coincidence of Aire and K14 expression by MTEC was evident in the Aire−/− thymus, while the coincidence of K14 and p63 expression was maintained (Figures 4 i & j).

Figure 4
The frequency of p63 MTEC is reduced in the Aire−/− thymus. Demonstration of p63 expression in Aire+/+ mice and Aire−/− mice with counterstaining to identify expression of Aire and stromal cell markers. The uppermost panels ...

Morphometric analyses of the frequency of p63+ MTEC in thymus tissue confirmed visual impressions, where the medullary density of these cells was reduced by about 50% in the Aire−/− thymus, with no discernable differences in cortical expression of p63 (Figure 5a). The reduced expression of p63 by Aire−/− MTEC was also evident in sorted MTEC populations as measured by quantitative PCR. This reduction primarily involved the expression of the truncated ΔN p63 isoforms; reduced expression of the full length TA isoforms of p63 did not appear to be significantly affected. Furthermore, there were comparable reductions in expression of the α, β, and γ DN isoforms (Table 1).

Figure 5
Reduced expression of p63 and p63 target genes in the Aire−/− thymus. a. Reduction of p63 expression in the Aire−/− thymus is restricted to the medulla. Multiple sections from 3 Aire+/+ and 3 Aire−/− mice ...
Table 1
P63 expression by isolated CD45, EpCam+ MTEc2

Reduced expression of p63 target genes in the Aire-deficient thymus

The reduction of p63 expression in the Aire-deficient thymus demonstrated by Immunohistochemistry and PCR analysis was indirectly confirmed by the demonstration of corresponding reduced expression of several p63 targets as judged by quantitative PCR. These included Perp (39), keratin 14 (15 , 40, 41), Edar, FGFR2IIIb, Jagged 1 and Jagged 2 (8, 10 , 16) (Figure 5b and data not shown). Expression of selected p63 targets that are positively regulated by TA p63 isoforms, such as IKKα, (42) or p21 (43) was minimally altered in the Aire−/− MTEC(data not shown).

Patterns of MTEC apoptosis in Aire+/+ and Aire−/− thymus

Due to the high levels of apoptosis exhibited by TE recovered by enzymatic digestion ( (19) Gillard and Farr, unpublished observations), a comparison of TE apoptosis in Aire+/+ and Aire−/− thymi required an in situ approach. We employed either a TUNEL assay or a fluorescent caspase inhibitor in combination with immunohistochemisty to characterize patterns of MTEC apoptosis in the Aire+/+ and Aire−/− thymus. As shown in Figure 6, both approaches yielded similar results, demonstrating that apoptotic cells in the Aire+/+ thymus were predominantly K8+K14+, while in the Aire−/− thymus, apoptotic MTEC were predominantly K8+K14−/low. Because the frequency of apoptotic cells in the medulla was fairly low, and only a subset of apoptotic cells could be clearly identified as keratin+ MTEC, the number of MTEC analyzed was too low to provide quantitative data. However, there was a clear shift in the phenotype of MTEC that displayed evidence of DNA fragmentation or caspase activation.

Figure 6
MTEC apoptosis in Aire+/+ and Aire−/− thymi. Perfusion-fixed thymic tissue from Aire+/+ (upper panel) and Aire−/− (lower panel) were processed to demonstrate the phenotype of apoptotic cells identified with either the Tunel ...


This study raises a number of important issues regarding models of TE differentiation. First, data presented here indicate that Aire-deficiency provides new insight into the terminal differentiation program of MTEC that is normally obscured by the apoptotic activity of Aire in the WT thymus. While it has been convincingly demonstrated that MTEC expressing Aire represent a post-mitotic population that turn over rapidly (19), the differentiation status of these cells is not known. This is an important issue because cessation of proliferation, widely considered to represent the Initiation of the terminal differentiation program, can be followed by extensive differentiation by postmitotic cells. For instance, progeny of the mitotically active epithelial cells in the basal layer of epidermis become post-mitotic as they enter the suprabasal spinous layer, and then undergo substantial subsequent differentiation as they progress through the granular layer to form the cornified layer, which represents completion of their terminal differentiation program (44) (shown diagrammatically in Figure 7a). Thus, the MTEC eliminated by Aire may be post-mitotic end-stage cells that have completed their program of terminal differentiation, or may be postmitotic cells that are eliminated before this program is completed. These two scenarios make very different predictions regarding the impact of Aire-deficiency on the composition of the medullary compartment. If the MTECs eliminated by Aire represent an end-stage, terminally differentiated population, loss of Aire activity in Aire−/− mice should lead to accumulation of MTEC that are indistinguishable from the end-stage Aire+ MTEC in Aire+/+ mice. This scenario also predicts that apoptotic MTEC in Aire+/+ and Aire−/− thymi would be indistinguishable because they would represent cells eliminated at the same end-stage of terminal differentiation (shown diagrammatically in Figure 7b). Alternatively, if Aire eliminates MTEC that are post-mitotic but have not completed their program of terminal differentiation, Aire−/− MTEC could complete this program, leading to the accumulation of terminally differentiated MTEC. In this case, subsets of MTEC that may normally represent minor populations of terminally differentiated MTEC in the Aire+/+ thymus would be expanded in the Aire−/− thymus (shown diagrammatically in Figure 7c).

Figure 7
Schematic representation of stages of epithelial differentiation and how Aire-mediated MTEC apoptosis could affect the pattern of MTEC differentiation. Different stages of epithelial differentiation are indicated by shapes. Panel a depicts the situation ...

Some of the alteration of the Aire−/− medullary compartment described here would be consistent with either interpretation; reduction of p63+ MTEC and an expansion of MTEC expressing truncated, non-functional Aire protein would be sequelae of enhanced survival of post-mitotic MTEC regardless of their stage of differentiation. However, other features of the Aire−/− MTEC are difficult to reconcile with Aire expression and apoptotic elimination targeting end-stage terminally differentiated MTEC, and would be consistent with an Aire-dependent interruption of a terminal differentiation process. One is the shift in Aire expression from K14+/high K8+/low cells in the Aire+/+ thymus toward K14−/low K8+ MTEC in the Aire−/− thymus. As will be discussed below, similar shifts in patterns of keratin expression reflect progressive differentiation in other epithelial tissues. This phenotypic change is accompanied by a shift in the morphology of the Aire+ MTEC from a more reticular pattern in the Aire+/+ medulla to more globular cells in the Aire−/− thymus. Furthermore, the predominant K14+K8+ phenotype of apoptotic MTEC in the Aire+/+ thymus is contrasted by the elevated representation of K14−/low apoptotic MTEC in the Aire−/− thymus, indicating that Aire+/+ and Aire−/− MTEC are eliminated at different stages of differentiation.

A working hypothesis based on these data is that end-stage, terminally differentiated MTEC acquire a K8+K14−/low phenotype and are represented by the small subset of globular K8+K4/K5−/low MTEC seen in the normal thymus (6, 7, 31). The minor representation of these cells in the normal medulla would reflect Aire-mediated elimination of many MTEC at a more proximal stage of terminal differentiation. Aire-deficiency abrogating this apoptotic interruption of the MTEC differentiation program would explain the accumulation of putative terminally differentiated globular K8+K14−/low terminally differentiated MTEC in the Aire−/− thymus (20).

Admittedly, this interpretation of MTEC differentiation runs counter to the prevailing view. However, several lines of evidence indicate that alternative models of TE differentiation based on the differentiation of other epithelial tissues should be explored. One concerns the role of p63. The impact of p63-deficiency on thymic epithelium indicates that p63 contributes to the maintenance and/or behavior of TE at proximal stages of their differentiation (8, 9). Although there are indications that p63 expression identifies progenitor epithelial cells (15), the extensive expression of p63 by thymic epithelial cells (3538); this report) indicates that p63 expression in the thymus is not restricted to this relatively rare population (4) and includes transit amplifying cells as well. There is debate regarding the role of p63 in epithelial differentiation, where it has been implicated in differentiation, cell fate specification, proliferation, survival, senescence, apoptosis, cell-cell and cell-matrix interaction programs (reviewed in (13, 16, 45)). In epidermal epithelium, p63 is expressed prominently by basal, mitotically active K14+ progenitor cells and is down-regulated in more differentiated, postmitotic keratinocytes that express lower levels of K14 and express Keratins 1 and 10 instead (14, 46). It has been recently demonstrated that down-regulation of p63 expression in basal epithelial cells by the action of a microRNA serves as a signal for initiation of keratinocyte terminal differentiation (47, 48). Thus, for purposes of discussion here, we have taken the view that p63 expression is associated with epithelial cells that are either proliferating or retain proliferative potential (14). The extensive expression of p63 by cortical and medullary thymic epithelium is consistent with recent reports indicating that TE undergoes considerable turnover (31, 4951), while the minimal Aire and p63 co-expression by K14+K8+ MTEC shown here would be consistent with the non-mitotic state proposed for Aire+ MTEC (19).

Data presented here also help refine the current model of TE differentiation based on patterns of keratin expression. CTEC and MTEC compartments have been widely identified on the basis of reciprocal expression of K8 and K14, respectively. Expression patterns of K5, a major pairing partner with K14, has a broader expression pattern among cortical TE and a strong circumstantial case has been made that K5+K8+ TE are the immediate precursors to K8+K5 CTEC (6, 7). Based on these initial reports, the MTEC population is widely considered to be largely K5+K8 and/or K14+K8, except for scattered globular K8+K5/K14 cells. The widespread, albeit variable expression of K8 by K5/K14+medullary TE (19, 20) this report), and the hypothesis put forward here that end-stage terminally differentiated MTEC acquire a K14K8+ phenotype, suggest an alternative to the previously proposed binary pattern of K5/K14 and K8 expression by MTEC. The program of prostatic epithelial (PE) differentiation serves as a useful comparison because this epithelium bears a number of similarities to TE. Basal PE cells express p63, high levels of K5 and K14, and low levels of K8, a phenotype expressed by many MTEC. These basal PE terminally differentiate to become luminal epithelium lacking p63, K5 or K14, while expressing high levels of K8/18, a phenotype expressed by many CTEC and a significant subset of MTEC in the Aire−/− medulla. A basal intermediate PE population, derived from p63+K5+K14+ cells, displays a p63+K5+K14K8+ phenotype (17, 5254) and thus resembles the TE proposed to be the immediate precursors to terminally differentiated CTEC (6, 7) (shown diagrammatically in Figure 8a). This model applied to thymic epithelial differentiation (Figure 8b) suggests that p63+K14+K5+K8+ MTEC population contain progenitor/transit amplifying cells, a suggestion consistent with the demonstration that the keratin14 promoter is active in a subset of postnatal epithelial cells able to give rise to a complete thymic epithelial compartment (4). The terminal differentiation program of these cells would involve coordinate loss of mitotic potential and p63 expression and down-regulation of K5 and K14 expression, while levels of K8 expression would be maintained or elevated. As a functional K14 enhancer is a direct positively regulated target of p63 (55), coordinate reduction of p63 and K14 expression is not surprising. Based on the postmitotic status of Aire+ cells shown previously (19) and the p63/keratin phenotype of Aire+ MTEC described here, it is suggested that initiation of MTEC terminal differentiation, down-regulation of p63 and cessation of mitotic activity, occurs in a subset of K14+K8+ MTEC. Due to extensive elimination of MTEC at this stage of differentiation, few MTEC complete the terminal differentiation program, accounting for the low frequency of p63K14K8+ MTEC in the normal medullary compartment. However, in the absence of Aire-mediated apoptosis, the p63K14+K8+ MTEC can complete down-regulation of K14 expression and acquire a K8+K14−/low phenotype and globular morphology before they also undergo apoptosis upon completion of their differentiation program. This model proposes a symmetry of CTEC and MTEC differentiation programs in that they would both originate from a K14+K5+ progenitor population and because terminal differentiation would lead to cells expressing high levels of K8 relative to K5/K14 in either compartment. Differences that emerge between CTEC and MTEC, such as the more precipitous down-regulation of K14 expression by CTEC (or persistent K5/K14 expression by MTEC) and other differences in programs of gene expression evidenced by CTEC and MTEC, may reflect the unique set of local signals experienced by TE in these different environments. MTEC likely experience a more complex environment in terms of cellular constituents and the milieu of locally produced mediators. For instance the CD4+CD3 lymphoid tissue-inducer cells that express RANKL and can regulate Aire expression, have a predominantly medullary localization (56), as do mature thymocytes that may modulate levels of Aire expression through CD40-CD40 ligand interactions (57). In this context, it is possible that the CD80+MHCIIhigh phenotype and expression of Aire by MTEC represent situational responses of these cells to local signals and may not be features intrinsic to the MTEC differentiation program This point is raised because a CD80+ MHCII+ phenotype can be readily induced by multiple cell types in response to proinflammatory cytokines (5861).

Figure 8
Schematic representation of epidermal differentiation occurring in prostatic and thymic epithelium. Patterns of keratin expression during the differentiation of cortical and medullary epithelium bear strong resemblance to the program displayed by prostatic ...

The predominance of p63+K14+K8+ transit amplifying MTEC in the Aire+/+ thymus and the emergence of a distinctive p63K14K8+ in the Aire−/− thymus indicates that the pro-apoptotic activity of Aire affects the relative abundance of MTEC at different stages of differentiation. This provides an explanation for the previous observation that expression of transcription factors associated with progenitor cells, Sox2, Oct4, and Nanog (62), could be detected in pools of sorted MTEC from Aire+/+ but not Aire−/− mice (20). While this may reflect their direct transcriptional control by Aire, these transcription factors may not be detected in sorted Aire−/− MTEC because the representation of immature MTEC that express these transcription factors has been reduced to the point where the levels of these transcription factors are below detection. Similarly, if K14+ MTEC represent a relatively immature population of cells as we suggest here, the reduced expression of K14+ Aire−/− by Aire MTEC (20), this report) would also be consistent with an expansion of mature MTEC at the expense of immature cells.

Finally, the proposal here that Aire expression occurs when MTEC initiate their program of terminal differentiation may be relevant to Aire-dependent regulation of tissue restricted antigens (TRA). If Aire act to de-repress gene expression by process involving chromatin remodeling and altered transcriptional activity, this process may be more prevalent in less differentiated cells that have only begun to make the chromatin modifications associated with terminal differentiation. It has also been proposed that Aire modifies RNA processing (63). Thus, Aire might regulate the “gain” or efficiency of transcription at a given locus, rather than altering patterns of transcription. From this perspective, transit amplifying MTEC initiating their terminal differentiation program might be expected to display less of the silencing/ inactivation of transcriptional networks that is a hallmark of the differentiation process and hence, have greater heterogeneity of transcriptional substrates for Aire-mediated amplification, compared to MTEC that are terminally differentiated (discussed in (20, 64)).

Definitive evaluation of the role of Aire in MTEC differentiation and TRA expression await approaches to fate-map the differentiation program of MTEC and a better understanding of the stages of TE differentiation. Data presented here indicate that Aire−/− MTEC provide a valuable starting point to define aspects of the terminal stages of this process.


We thank Dr. L. Peltonen for providing Aire−/− mice and Drs. A. Liston, A. Rudensky, and B. Kyewski for comments, suggestions, criticisms, and challenging discussions.


1This work was supported by Grant Number AI09575 from the NIAID of the NIH. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIAID.


The authors have no financial conflict of interest.


1. Hollander G, Gill J, Zuklys S, Iwanami N, Liu C, Takahama Y. Cellular and molecular events during early thymus development. Immunol Rev. 2006;209:28–46. [PubMed]
2. Anderson G, Lane PJ, Jenkinson EJ. Generating intrathymic microenvironments to establish T-cell tolerance. Nat Rev Immunol. 2007;7:954–963. [PubMed]
3. Rossi SW, Jenkinson WE, Anderson G, Jenkinson EJ. Clonal analysis reveals a common progenitor for thymic cortical and medullary epithelium. Nature. 2006;441:988–991. [PubMed]
4. Bleul CC, Corbeaux T, Reuter A, Fisch P, Monting JS, Boehm T. Formation of a functional thymus initiated by a postnatal epithelial progenitor cell. Nature. 2006;441:992–996. [PubMed]
5. Rossi SW, Chidgey AP, Parnell SM, Jenkinson WE, Scott HS, Boyd RL, Jenkinson EJ, Anderson G. Redefining epithelial progenitor potential in the developing thymus. European journal of immunology. 2007;37:2411–2418. [PubMed]
6. Klug DB, Carter C, Crouch E, Roop D, Conti CJ, Richie ER. Interdependence of cortical thymic epithelial cell differentiation and T-lineage commitment. Proceedings of the National Academy of Sciences of the United States of America. 1998;95:11822–11827. [PMC free article] [PubMed]
7. Klug DB, Carter C, Gimenez-Conti IB, Richie ER. Cutting edge: thymocyte-independent and thymocyte-dependent phases of epithelial patterning in the fetal thymus. J Immunol. 2002;169:2842–2845. [PubMed]
8. Candi E, Rufini A, Terrinoni A, Giamboi-Miraglia A, Lena AM, Mantovani R, Knight R, Melino G. DeltaNp63 regulates thymic development through enhanced expression of FgfR2 and Jag2. Proceedings of the National Academy of Sciences of the United States of America. 2007;104:11999–12004. [PMC free article] [PubMed]
9. Senoo M, Pinto F, Crum CP, McKeon F. p63 Is essential for the proliferative potential of stem cells in stratified epithelia. Cell. 2007;129:523–536. [PubMed]
10. Laurikkala J, Mikkola ML, James M, Tummers M, Mills AA, Thesleff I. p63 regulates multiple signalling pathways required for ectodermal organogenesis and differentiation. Development. 2006;133:1553–1563. [PubMed]
11. Revest JM, Suniara RK, Kerr K, Owen JJ, Dickson C. Development of the thymus requires signaling through the fibroblast growth factor receptor R2-IIIb. J Immunol. 2001;167:1954–1961. [PubMed]
12. Dooley J, Erickson M, Larochelle WJ, Gillard GO, Farr AG. FGFR2IIIb signaling regulates thymic epithelial differentiation. Dev Dyn. 2007;236:3459–3471. [PubMed]
13. Candi E, Dinsdale D, Rufini A, Salomoni P, Knight RA, Mueller M, Krammer PH, Melino G. TAp63 and DeltaNp63 in cancer and epidermal development. Cell Cycle. 2007;6:274–285. [PubMed]
14. Parsa R, Yang A, McKeon F, Green H. Association of p63 with proliferative potential in normal and neoplastic human keratinocytes. J Invest Dermatol. 1999;113:1099–1105. [PubMed]
15. Pellegrini G, Dellambra E, Golisano O, Martinelli E, Fantozzi I, Bondanza S, Ponzin D, McKeon F, De Luca M. p63 identifies keratinocyte stem cells. Proceedings of the National Academy of Sciences of the United States of America. 2001;98:3156–3161. [PMC free article] [PubMed]
16. Perez CA, Pietenpol JA. Transcriptional programs regulated by p63 in normal epithelium and tumors. Cell Cycle. 2007;6:246–254. [PubMed]
17. Isaacs JT, Coffey DS. Etiology and disease process of benign prostatic hyperplasia. The Prostate. 1989;2:33–50. [PubMed]
18. Derbinski J, Schulte A, Kyewski B, Klein L. Promiscuous gene expression in medullary thymic epithelial cells mirrors the peripheral self. Nat Immunol. 2001;2:1032–1039. [PubMed]
19. Gray D, Abramson J, Benoist C, Mathis D. Proliferative arrest and rapid turnover of thymic epithelial cells expressing Aire. J Exp Med. 2007;204:2521–2528. [PMC free article] [PubMed]
20. Gillard GO, Dooley J, Erickson M, Peltonen L, Farr AG. Aire-dependent alterations in medullary thymic epithelium indicate a role for Aire in thymic epithelial differentiation. J Immunol. 2007;178:3007–3015. [PubMed]
21. Anderson MS, Venanzi ES, Klein L, Chen Z, Berzins SP, Turley SJ, von Boehmer H, Bronson R, Dierich A, Benoist C, Mathis D. Projection of an immunological self shadow within the thymus by the aire protein. Science. 2002;298:1395–1401. [PubMed]
22. Ramsey C, Winqvist O, Puhakka L, Halonen M, Moro A, Kampe O, Eskelin P, Pelto-Huikko M, Peltonen L. Aire deficient mice develop multiple features of APE ED phenotype and show altered immune response. Hum Mol Genet. 2002;11:397–409. [PubMed]
23. Farr A, Nelson A, Truex J, Hosier S. Epithelial heterogeneity in the murine thymus: a cell surface glycoprotein expressed by subcapsular and medullary epithelium. J Histochem Cytochem. 1991;39:645–653. [PubMed]
24. Farr A, Nelson A, Hosier S. Characterization of an antigenic determinant preferentially expressed by type I epithelial cells in the murine thymus. J Histochem Cytochem. 1992;40:651–664. [PubMed]
25. Farr A, Nelson A, Hosier S, Kim A. A novel cytokine-responsive cell surface glycoprotein defines a subset of medullary thymic epithelium in situ. J Immunol. 1993;150:1160–1171. [PubMed]
26. Dooley J, Erickson M, Farr AG. An organized medullary epithelial structure in the normal thymus expresses molecules of respiratory epithelium and resembles the epithelial thymic rudiment of nude mice. J Immunol. 2005;175:4331–4337. [PubMed]
27. Kraal G, Breel M, Janse M, Bruin G. Langerhans' cells, veiled cells, and interdigitating cells in the mouse recognized by a monoclonal antibody. J Exp Med. 1986;163:981–997. [PMC free article] [PubMed]
28. Rouse RV, Bolin LM, Bender JR, Kyewski BA. Monoclonal antibodies reactive with subsets of mouse and human thymic epithelial cells. J Histochem Cytochem. 1988;36:1511–1517. [PubMed]
29. Brulet P, Babinet C, Kemler R, Jacob F. Monoclonal antibodies against trophectoderm-specific markers during mouse blastocyst formation. Proceedings of the National Academy of Sciences of the United States of America. 1980;77:4113–4117. [PMC free article] [PubMed]
30. Surh CD, Sprent J. T-cell apoptosis detected in situ during positive and negative selection in the thymus. Nature. 1994;372:100–103. [PubMed]
31. Gillard GO, Farr AG. Features of medullary thymic epithelium implicate postnatal development in maintaining epithelial heterogeneity and tissue-restricted antigen expression. J Immunol. 2006;176:5815–5824. [PubMed]
32. Heino M, Peterson P, Kudoh J, Nagamine K, Lagerstedt A, Ovod V, Ranki A, Rantala I, Nieminen M, Tuukkanen J, Scott HS, Antonarakis SE, Shimizu N, Krohn K. Autoimmune regulator is expressed in the cells regulating immune tolerance in thymus medulla. Biochem Biophys Res Commun. 1999;257:821–825. [PubMed]
33. Heino M, Peterson P, Sillanpaa N, Guerin S, Wu L, Anderson G, Scott HS, Antonarakis SE, Kudoh J, Shimizu N, Jenkinson EJ, Naquet P, Krohn KJ. RNA and protein expression of the murine autoimmune regulator gene (Aire) in normal, RelB-deficient and in NOD mouse. European journal of immunology. 2000;30:1884–1893. [PubMed]
34. Hubert FX, Kinkel SA, Webster KE, Cannon P, Crewther PE, Proeitto AI, Wu L, Heath WR, Scott HS. A specific anti-aire antibody reveals aire expression is restricted to medullary thymic epithelial cells and not expressed in periphery. J Immunol. 2008;180:3824–3832. [PubMed]
35. Dotto J, Pelosi G, Rosai J. Expression of p63 in thymomas and normal thymus. Am J Clin Pathol. 2007;127:415–420. [PubMed]
36. Chilosi M, Zamo A, Brighenti A, Malpeli G, Montagna L, Piccoli P, Pedron S, Lestani M, Inghirami G, Scarpa A, Doglioni C, Menestrina F. Constitutive expression of DeltaN-p63alpha isoform in human thymus and thymic epithelial tumours. Virchows Arch. 2003;443:175–183. [PubMed]
37. Kikuchi T, Ichimiya S, Kojima T, Crisa L, Koshiba S, Tonooka A, Kondo N, Van Der Saag PT, Yokoyama S, Sato N. Expression profiles and functional implications of p53-like transcription factors in thymic epithelial cell subtypes. Int Immunol. 2004;16:831–841. [PubMed]
38. Irifune T, Tamechika M, Adachi Y, Tokuda N, Sawada T, Fukumoto T. Morphological and immunohistochemical changes to thymic epithelial cells in the irradiated and recovering rat thymus. Arch Histol Cytol. 2004;67:149–158. [PubMed]
39. Ihrie RA, Marques MR, Nguyen BT, Horner JS, Papazoglu C, Bronson RT, Mills AA, Attardi LD. Perp is a p63-regulated gene essential for epithelial integrity. Cell. 2005;120:843–856. [PubMed]
40. Koster MI, Kim S, DeMayo FJ, Roop DR. p63 is the molecular switch for initiation of an epithelial stratification program. Genes & development. 2004;18:126–131. [PMC free article] [PubMed]
41. Candi E, Rufini A, Terrinoni A, Dinsdale D, Ranalli M, Paradisi A, De Laurenzi V, Spagnoli LG, Catani MV, Ramadan S, Knight RA, Melino G. Differential roles of p63 isoforms in epidermal development: selective genetic complementation in p63 null mice. Cell death and differentiation. 2006;13:1037–1047. [PubMed]
42. Candi E, Terrinoni A, Rufini A, Chikh A, Lena AM, Suzuki Y, Sayan BS, Knight RA, Melino G. p63 is upstream of IKK alpha in epidermal development. J Cell Sci. 2006;119:4617–4622. [PubMed]
43. Westfall MD, Mays DJ, Sniezek JC, Pietenpol JA. The Delta Np63 alpha phosphoprotein binds the p21 and 14-3-3 sigma promoters in vivo and has transcriptional repressor activity that is reduced by Hay-Wells syndromederived mutations. Molecular and cellular biology. 2003;23:2264–2276. [PMC free article] [PubMed]
44. Nagarajan P, Romano RA, Sinha S. Transcriptional control of the differentiation program of interfollicular epidermal keratinocytes. Critical reviews in eukaryotic gene expression. 2008;18:57–79. [PubMed]
45. McKeon F. p63 and the epithelial stem cell: more than status quo? Genes & development. 2004;18:465–469. [PubMed]
46. Yang A, Kaghad M, Wang Y, Gillett E, Fleming MD, Dotsch V, Andrews NC, Caput D, McKeon F. p63, a p53 homolog at 3q27-29, encodes multiple products with transactivating, death-inducing, and dominantnegative activities. Mol Cell. 1998;2:305–316. [PubMed]
47. Lena AM, Shalom-Feuerstein R, di Val Cervo PR, Aberdam D, Knight RA, Melino G, Candi E. miR-203 represses 'stemness' by repressing DeltaNp63. Cell death and differentiation. 2008 [PubMed]
48. Yi R, Poy MN, Stoffel M, Fuchs E. A skin microRNA promotes differentiation by repressing 'stemness'. Nature. 2008;452:225–229. [PubMed]
49. Yang SJ, Ahn S, Park CS, Holmes KL, Westrup J, Chang CH, Kim MG. The quantitative assessment of MHC II on thymic epithelium: implications in cortical thymocyte development. Int Immunol. 2006;18:729–739. [PubMed]
50. Gray DH, Seach N, Ueno T, Milton MK, Liston A, Lew AM, Goodnow CC, Boyd RL. Developmental kinetics, turnover, and stimulatory capacity of thymic epithelial cells. Blood. 2006;108:3777–3785. [PubMed]
51. Rossi SW, Jeker LT, Ueno T, Kuse S, Keller MP, Zuklys S, Gudkov AV, Takahama Y, Krenger W, Blazar BR, Hollander GA. Keratinocyte growth factor (KGF) enhances postnatal T-cell development via enhancements in proliferation and function of thymic epithelial cells. Blood. 2007;109:3803–3811. [PMC free article] [PubMed]
52. Hudson DL, O'Hare M, Watt FM, Masters JR. Proliferative heterogeneity in the human prostate: evidence for epithetlial stem cells. Laboratory investigation; a journal of technical methods and pathology. 2000;80:1243–1250. [PubMed]
53. Robinson EJ, Neal DE, Collins AT. Basal cells are progenitors of luminal cells in primary cultures of differentiating human prostatic epithelium. Prostate. 1998;37:149–160. [PubMed]
54. van Leenders G, Dijkman H, Hulsbergen-van de Kaa C, Ruiter D, Schalken J. Demonstration of intermediate cells during human prostate epithelial differentiation in situ and in vitro using triple-staining confocal scanning microscopy. Laboratory investigation; a journal of technical methods and pathology. 2000;80:1251–1258. [PubMed]
55. Romano RA, Birkaya B, Sinha S. A functional enhancer of keratin14 is a direct transcriptional target of deltaNp63. J Invest Dermatol. 2007;127:1175–1186. [PubMed]
56. Rossi SW, Kim MY, Leibbrandt A, Parnell SM, Jenkinson WE, Glanville SH, McConnell FM, Scott HS, Penninger JM, Jenkinson EJ, Lane PJ, Anderson G. RANK signals from CD4(+)3(−) inducer cells regulate development of Aire-expressing epithelial cells in the thymic medulla. J Exp Med. 2007;204:1267–1272. [PMC free article] [PubMed]
57. White AJ, Withers DR, Parnell SM, Scott HS, Finke D, Lane PJ, Jenkinson EJ, Anderson G. Sequential phases in the development of Aire-expressing medullary thymic epithelial cells involve distinct cellular input. European journal of immunology. 2008;38:942–947. [PubMed]
58. Battifora M, Pesce G, Paolieri F, Fiorino N, Giordano C, Riccio AM, Torre G, Olive D, Bagnasco M. B7.1 costimulatory molecule is expressed on thyroid follicular cells in Hashimoto's thyroiditis, but not in Graves' disease. The Journal of clinical endocrinology and metabolism. 1998;83:4130–4139. [PubMed]
59. Elssner A, Jaumann F, Wolf WP, Schwaiblmair M, Behr J, Furst H, Reichenspurner H, Briegel J, Niedermeyer J, Vogelmeier C. Bronchial epithelial cell B7-1 and B7-2 mRNA expression after lung transplantation: a role in allograft rejection? Eur Respir J. 2002;20:165–169. [PubMed]
60. Matsumura R, Umemiya K, Goto T, Nakazawa T, Kagami M, Tomioka H, Tanabe E, Sugiyama T, Sueishi M. Glandular and extraglandular expression of costimulatory molecules in patients with Sjogren's syndrome. Annals of the rheumatic diseases. 2001;60:473–482. [PMC free article] [PubMed]
61. Nikcevich KM, Gordon KB, Tan L, Hurst SD, Kroepfl JF, Gardinier M, Barrett TA, Miller SD. IFN-gamma-activated primary murine astrocytes express B7 costimulatory molecules and prime naive antigen-specific T cells. J Immunol. 1997;158:614–621. [PubMed]
62. Boyer LA, Lee TI, Cole MF, Johnstone SE, Levine SS, Zucker JP, Guenther MG, Kumar RM, Murray HL, Jenner RG, Gifford DK, Melton DA, Jaenisch R, Young RA. Core transcriptional regulatory circuitry in human embryonic stem cells. Cell. 2005;122:947–956. [PMC free article] [PubMed]
63. Oven I, Brdickova N, Kohoutek J, Vaupotic T, Narat M, Peterlin BM. AIRE recruits P-TEFb for transcriptional elongation of target genes in medullary thymic epithelial cells. Molecular and cellular biology. 2007;27:8815–8823. [PMC free article] [PubMed]
64. Gillard GO, Farr AG. Contrasting models of promiscuous gene expression by thymic epithelium. J Exp Med. 2005;202:15–19. [PMC free article] [PubMed]
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...